Evaluation of Grey Area Mitigation Tools within Zonal and Non-Zonal RANS-LES Approaches in Flows with Pressure Induced Separation

Abstract

A study is presented evaluating the performance of zonal and non-zonal hybrid RANS-LES approaches for a flow with pressure induced separation with special emphasis of the alleviation of Grey Area effects. The test case chosen for the evaluation is the flow about the so-called 2D wall-mounted hump, which was designed as a CFD validation test case. In the study two zonal RANS-WMLES approaches employing Improved Delayed DES (IDDES) are compared with one non-zonal hybrid approach using Delayed DES (IDDES) with a modified Shear Layer Adapted definition of the subgrid length-scale. All approaches use k- SST as the underlying RANS turbulence model, but the two zonal methods employ two quite different flow solvers each with its own method to inject turbulent content at the RANS-LES interface. These solvers are the unstructured compressible DLR TAU code employing the Synthetic Eddy Method and the higher order structured code of NTS, which is used in its incompressible mode with the Synthetic Turbulence Generator of NTS. With both codes solutions are obtained with the zonal approach varying the position of the interface between RANS and LES to study not only the influence of the different methods but also the influence of the location of turbulence injection on the simulation quality. The corresponding results and those from the additional non-zonal simulation with the NTS code are compared against each other and against available experimental data, comprising wall pressure and skin-friction distributions as well as mean velocity and Reynolds stresses profiles. Overall very satisfactory performance of the approaches employed as well as good agreement with the experiment will be presented.